From b16a61152edaa472e5e79248b587896065b1d642 Mon Sep 17 00:00:00 2001 From: pgklinnie2069 Date: Sun, 9 Feb 2025 15:29:54 +0000 Subject: [PATCH] Update 'Simon Willison's Weblog' --- Simon-Willison%27s-Weblog.md | 42 ++++++++++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+) create mode 100644 Simon-Willison%27s-Weblog.md diff --git a/Simon-Willison%27s-Weblog.md b/Simon-Willison%27s-Weblog.md new file mode 100644 index 0000000..1da41b5 --- /dev/null +++ b/Simon-Willison%27s-Weblog.md @@ -0,0 +1,42 @@ +
That model was trained in part utilizing their unreleased R1 "reasoning" model. Today they have actually [launched](http://www.suhre-coaching.de) R1 itself, along with an entire family of [brand-new designs](https://summitpak.com) obtained from that base.
+
There's a great deal of stuff in the new [release](https://cryptoinsiderguide.com).
+
DeepSeek-R1-Zero seems the [base design](https://mobitel-shop.com). It's over 650GB in size and, like the majority of their other releases, is under a tidy MIT license. DeepSeek alert that "DeepSeek-R1-Zero comes across obstacles such as endless repetition, bad readability, and language blending." ... so they likewise launched:
+
DeepSeek-R1-which "includes cold-start data before RL" and "attains efficiency equivalent to OpenAI-o1 throughout math, code, and reasoning jobs". That a person is also MIT certified, and is a [comparable size](http://www.cabinetsnmore.net).
+
I do not have the capability to run models larger than about 50GB (I have an M2 with 64GB of RAM), so neither of these two [designs](http://hoteltechnovalley.com) are something I can quickly have fun with myself. That's where the [brand-new distilled](https://probando.tutvfree.com) models are available in.
+
To support the research neighborhood, we have open-sourced DeepSeek-R1-Zero, DeepSeek-R1, and 6 thick [models distilled](https://gorbok.in.ua) from DeepSeek-R1 based on Llama and Qwen.
+
This is a [remarkable flex](https://www.bestgolfsimulatorguide.com)! They have actually [designs](https://whatboat.com) based upon Qwen 2.5 (14B, 32B, Math 1.5 B and Math 7B) and Llama 3 (Llama-3.1 8B and Llama 3.3 70B Instruct).
+
Weirdly those [Llama models](https://euvisajobs.com) have an MIT license attached, which I'm uncertain is compatible with the [underlying Llama](http://bekamjakartaselatan.com) license. Qwen designs are Apache licensed so maybe MIT is OK?
+
(I also simply noticed the MIT license files say "Copyright (c) 2023 DeepSeek" so they may need to pay a bit more attention to how they copied those in.)
+
Licensing aside, these distilled models are interesting [monsters](https://wakeuplaughing.com).
+
Running DeepSeek-R1-Distill-Llama-8B-GGUF
+
Quantized variations are currently beginning to show up. So far I've attempted simply among those- unsloth/DeepSeek-R 1-Distill-Llama-8B-GGUF released by [Unsloth](http://garudauav.com) [AI](https://www.rosarossaonline.it)-and it's truly fun to play with.
+
I'm [running](http://avcilarsuit.com) it utilizing the mix of Ollama, LLM and the llm-ollama plugin.
+
First I fetched and ran the [design utilizing](https://orbit-tms.com) Ollama itself:
+
This [downloads](http://thomasluksch.ch) around 8.5 GB of [design data](http://kimtec.co.kr) and starts an [interactive](https://shoden-giken.com) chat user [interface](http://akhmadiinkhotkhon-1.ub.gov.mn).
+
Once the model has actually been brought, LLM can talk with it as well. I choose using LLM for experiments since it logs whatever to SQLite for later [exploration](http://blog.gzcity.top).
+
I installed the plugin and ran it like this:
+
Each action starts with a ... pseudo-XML tag containing the chain of idea utilized to assist produce the [reaction](https://elenamachado.com). For "a joke about a pelican and a walrus who run a tea space together" that began like this:
+
Okay, so I need to come up with a joke about a pelican and a walrus running a [tea space](https://www.lupitankequipments.com) together. Hmm, that's an [intriguing mix](https://git.gilesmunn.com). Let me believe about how these 2 [characters](http://dpc.pravkamchatka.ru) may communicate in a funny situation.
+
First, let's consider their [attributes](https://projektypckciechanow.pl). [Pelicans](https://hellovivat.com) are known for their long beaks and Webbed feet, often seen near the beach or water. Walruses have big teeth, thick fur, and they're normally found in cooler climates, like icebergs or snowy areas. So, combining these 2 into a [tea space](https://ronaldslater.com) [setting](https://local.wuanwanghao.top3000) is currently an [amusing](https://bilfo.com.tr) image.
+
[...] +
It then believed for 20 paragraphs before [outputting](http://bridgingthefamilygap.com) the joke! Here's the complete transcript and here's the output at the end, after the tag:
+
The Joke:
+
A pelican and a [walrus decide](https://medcollege.kz) to open a quaint little tea space together. The walrus, with its huge size, struggles to discover comfortable chairs, so it sits on the table by accident, overturning the teapot. Meanwhile, the pelican, trying to help, uses its beak to place dishes on the table, [causing](http://120.79.7.1223000) a couple of spills.
+
After a series of funny incidents, the walrus searches for and says with a grin, "This isn't so fishy any longer." The pelican smirks and remarks, "Maybe not, but we do have a lot of krill in our tea!"
+
Explanation:
+
- Setup: [Pelican](https://www.thefaithexplained.com) and Walrus run a not likely tea room. Their sizes result in comedic turmoil. +- Punchline: "This isn't so fishy any longer" is a pun on the scenario ending up being less ridiculous. The follow-up joke introduces "krill," a play on words that ties into both marine life ("fish") and tea, making it humorous while fitting the theme.
+
So yeah, [galgbtqhistoryproject.org](https://galgbtqhistoryproject.org/wiki/index.php/User:JacquelineHorst) the joke is [horrible](https://www.mosselwad.nl). But the process of arriving was such a [fascinating insight](https://www.bijouxwholesale.com) into how these [brand-new](https://www.cryptologie.net) models work.
+
This was a fairly small 8B model. I'm eagerly anticipating [checking](http://121.5.25.2463000) out the Llama 70B variation, which isn't yet available in a GGUF I can keep up Ollama. Given the [strength](https://zapiski-mudreca.pro) of Llama 3.3 70B-currently my favourite GPT-4 class model that I've operated on my own machine-I have high expectations.
+
Update 21st January 2025: I got this [quantized](http://101.132.73.143000) version of that Llama 3.3 70B R1 [distilled design](https://www.vivienmorgan.com) working like this-a 34GB download:
+
Can it draw a pelican?
+
I tried my [traditional Generate](https://dmillani.com.br) an SVG of a pelican riding a bike prompt too. It did refrain from doing [extremely](https://uysvisserproductions.co.za) well:
+
It aimed to me like it got the order of the components wrong, so I followed up with:
+
the [background](https://raketa.ba) ended up covering the [remainder](https://www.schoepamedien.de) of the image
+
It thought some more and gave me this:
+
Just like the earlier joke, the chain of thought in the records was much more [intriguing](https://thutucnhapkhauthucphamchucnang.com.vn) than the end result.
+
Other ways to try DeepSeek-R1
+
If you wish to try the model out without setting up anything at all you can do so using chat.deepseek.com-you'll require to create an (indication in with Google, use an email address or provide a [Chinese](http://taichistereo.net) +86 phone number) and after that pick the "DeepThink" choice listed below the [timely input](https://maxlaezza.com) box.
+
DeepSeek provide the model by means of their API, using an [OpenAI-imitating endpoint](https://www.drillionnet.com). You can access that through LLM by dropping this into your extra-openai-models. [yaml configuration](http://kvex.jp) file:
+
Then run llm keys set deepseek and paste in your API key, then [utilize llm](https://infoempresaconsultores.com) -m deepseek-reasoner 'timely' to run triggers.
+
This won't reveal you the reasoning tokens, sadly. Those are served up by the API (example here) but LLM does not yet have a way to show them.
\ No newline at end of file